Pain associated with knee osteoarthritis (OA) is characterized by (a) high levels of variability among individuals; and (b) modest associations with radiographic measures of tissue damage. Ethnic background is one important factor that contributes to variation in pain reports and pain-related behavior. Recent studies indicate that African-American, compared to white, individuals exhibit lower pain tolerance levels and relatively impaired pain regulatory mechanisms (e.g., blood pressure responses to stressors). There also is evidence that endogenous opioid release contributes to the function of these mechanisms. The overall aim of the proposed study, then, is detemine whether changes in opioid neurotransmission evoked by thermal heat stimulation partially mediate ethnic differences in thermal pain responses among patients with knee OA. Thirty-six men (18 African-American, 18 white) with knee OA will undergo PET bain imaging with the opiodergic radioligand [18F] fluoroethyl-diprenorphine under two conditions: (a) exposure to thermal heat stimulation, tailored to individual pain thresholds, that will produce similar, moderate, levels of thermal pain intensity across patients; and (b) exposure to thermal heat stimulation that will produce perceptions of warmth across patients (sensory control condition). Patients will produce visual analogue scale (VAS) ratings of pain intensity and unpleasantness following stimulation. We expect that African-American, compared to white patients, will (a) produce significantly greater increases in pain unpleasantness ratings; and (b) exhibit lower opioid neurotransmission in medial pain system structures (e.g., amygdala, anterior cingulate cortex) from sensory control to painful thermal heat stimulation, even after controlling for psychosocial variables such as depressive symptom levels. We also expect that stimulation-evoked change in opioid neurotransmission in medial pain system structures will partially mediate the ethnic group difference in change in pain unpleasantness ratings. We anticipate that the proposed research will generate new sttudies that will lead to (a) improved understanding of additional physiologic or psychosocial variables that contribute to ethnic differences in pain responses and endogenous pain regulatory function; (b) development of new or refinement of current pharmacologic or behavioral/psychosocial interventions for pain management; and (c) reductions in ethnic disparities in patients' expectations of and preferences for these interventions.